What is retail marketing that is expanding with AI?

In the domestic retail industry, an increasing number of companies are adopting the latest technologies such as AI and IoT as a strategy to further expand sales profits. In the future, we held a seminar for retailers who want to transform their own store operations into technology-based store operations, asking them to consider "Image Recognition AI" as an option.

This time, Mr. Yuichi Sasaki, Board Director and CTO of Neural Pocket Co., Ltd., which independently develops and owns AI technology that uses images, took the stage and talked about various examples of AI using images in the retail industry. rice field.

Teacher

Neural Pocket Co., Ltd.

Board Director CTO

Yuichi Sasaki

 

Biography: Graduated from the Department of Physics, Faculty of Science, University of Tokyo, Doctor of Science.
Engaged in research on black holes and supersymmetric particles at the European Organization for Nuclear Research (CERN) in Switzerland.
He analyzed big data and developed AI, and contributed to the discovery of the Higgs boson, which later won the Nobel Prize.
Participated in Neural Pocket from the desire to return the world's highest level of AI research knowledge to society.
We are engaged in a wide range of research on the technical elements necessary for AI to deliver valuable information, such as the development of image recognition AI, the development of infrastructure to process the obtained data, and the development of devices that lead to changes in people's behavior.

　

　

Neural Pocket aims to "build a society where data circulates" through AI technology and infrastructure development, and Mr. Sasaki said that there are two steps to improving marketing through data analysis. First of all, it is important to know what kind of data and how to "acquire data" by making full use of technology, and "analyze and utilize" that data to make use of it in business and marketing.

　

Based on these two steps, five case studies were introduced, and among them, AI analyzed a large number of photos posted by users with high fashion sensitivity, about 1 million per year, and based on that big data, scientific research was conducted. It was very interesting to see an example that leads to product development.

In addition, by utilizing technology that can measure the number of people and the time spent in a designated area, it is possible to acquire "customer data before purchase" that was difficult to obtain at stores, and to detect when the parking lot is full and provide customer feedback. In addition to providing smooth shopping, technology that can acquire customer information such as the frequency of returning to the store from the license plate, signage that can acquire customer information, the number of passers-by, and viewing conditions in addition to displaying advertisements. We have explained in detail the examples that can be used in the scene.

What are the necessary steps to realize image analysis AI technology in actual stores?

Even if you want to implement the technology introduced above at your own store, many people don't know where to start. Therefore, Mr. Nomoto from Macnica explained the configuration of image analysis AI and the preparation and knowledge necessary to realize it.

The mechanism for image analysis using AI consists of three steps: obtaining information from the image (input), AI analyzing the image, and outputting the results. In the process, AI is realized through the “learning” phase where AI learns and the “inference” phase where AI makes decisions.

Each phase requires a variety of hardware and software, one of which is the GPU, which excels at high-speed image processing.

For AI using images, "edge computing" will be introduced, in which the GPU is placed closer to the camera (at the edge). The benefits include reducing the burden on the network, reducing total processing time, and strengthening security. Macnica is capable of coordinating the realization of this kind of AI, so if you are interested, please feel free to contact us.

A session that even AI beginners can understand

On the day of the event, we organized the following agenda and introduced what can be achieved with image recognition AI. In the questionnaire after the seminar, we received comments such as "The specific examples were helpful" and "It was very easy to understand." We are planning the second seminar for retailers, so we hope you will be able to join us next time.

You can watch on-demand videos with a simple registration.

If you register from the form below, you can watch the seminar video on the day. If you missed it or if you are a participant and want to see it again, please register for this opportunity!